---
title: FislacBot
emoji: 📊
colorFrom: blue
colorTo: green
sdk: gradio
sdk_version: 5.6.0
app_file: app.py
pinned: false
license: mit
app_port: 7860
docker_user: root
base_path: /app
accelerator: gpu
---

FislacBot - AI Assistant for FISLAC Documentation
FislacBot is an artificial intelligence assistant specialized in FISLAC (Fiscal Latin America and Caribbean) documentation and fiscal analysis. It uses the Llama-2-7b model with RAG (Retrieval Augmented Generation) to provide accurate responses based on official documentation.
Author
Camilo Vega Barbosa

AI Professor and Artificial Intelligence Solutions Consultant
Connect with me:

LinkedIn
GitHub



Features

RAG-powered responses using official FISLAC documentation
Interactive chat interface using Gradio
GPU-accelerated inference
Context-aware responses with source tracking

How It Works
The application uses a sophisticated RAG system that:

Processes and indexes FISLAC documentation
Generates embeddings using multilingual-e5-large
Uses FAISS for efficient vector storage and retrieval
Combines retrieved context with Llama-2 for accurate responses

Technical Details

Model: Meta-llama/Llama-2-7b-chat-hf
Embeddings: intfloat/multilingual-e5-large
Vector Store: FAISS
Framework: Gradio
Dependencies: Managed through requirements.txt
Device Configuration: GPU-optimized using Accelerate

Installation
To run this application locally:

Clone the repository
Install dependencies:
bashCopypip install -r requirements.txt

Run the application:
bashCopypython app.py


Knowledge Base
The system is trained on:

Official FISLAC documentation
Valencia et al. (2022) - "Assessing macro-fiscal risk for Latin American and Caribbean countries"
Additional BID fiscal documentation


Created by Camilo Vega Barbosa, AI Professor and Solutions Consultant. For more AI projects and collaborations, feel free to connect on LinkedIn or visit my GitHub.